D2.2.Paper with Input from End Users.pdf

(756 KB) Pobierz
FP7-­‐SEC-­‐2011-­‐284725  
 
SURVEILLE  
Surveillance:  Ethical  Issues,  Legal  Limitations,  and  Efficiency    
Collaborative  Project  
 
SURVEILLE  Deliverable  2.2:  Paper  with  Input  from  End  Users  
 
Due  date  of  deliverable:  28.02.2013  
Actual  submission  date:  28.02.2013  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Start  date  of  project:   1.2.2012  
 
 
Duration:  39  months  
 
SURVEILLE  Work  Package  number  and  lead:  WP02  Prof.  Tom  Sorell  
Author(s):  Dr.  John  Guelke  
 
SURVEILLE:  Project  co-­‐funded  by  the  European  Commission  within  the  Seventh  Framework  Programme  
PU  
PP  
RE  
CO  
 
Public  
Dissemination  Level    
Restricted  to  other  programme  participants  (including  the  Commission  Services)  
Restricted  to  a  group  specified  by  the  consortium  (including  the  Commission  Services)  
Commission  
or  members  of  the  consortium  (including  the  Commission  Services)  
Confidential,  only  f
Services)  
X  
 
 
 
Executive  Summary  
 
1. The  earlier  FP7  project  DETECTER  project  constructed  a  normative  
framework  for  considering  the  ethical  risks  of  surveillance  technologies  in  
counter-­‐terrorism  investigations.  
2. This  is  compared  with  a  new  framework  devised  for  normative  assessments  
extracted  from  submissions  by  the  SURVEILLE  End  User  Panel  of  45  
surveillance  technology  products  presented  in  SURVEILLE  deliverable  D2.1.    
3. Although  there  is  substantial  overlap,  ethical  risks  of  surveillance  in  
SURVEILLE  arise  from  a  wider  range  of  situations  than  terrorism  which  was  in  
focus  for  DETECTER.  
4. The  normative  grounding  for  ethical  risk  is  considered  in  relation  to  five  
possible  features  of  serious  crime:  significant  financial  loss  to  the  victim,  use  
of  violence,  threat  to  public  order,  organisation,  and  significant  financial  gain  
for  the  perpetrator.  
 
 
1.  The  DETECTER  Normative  Framework  
 
The  DETECTER  project
1
 analysed  the  ethical  and  legal  norms  of  the  use  of  detection  
technology  in  counter-­‐terrorism  investigations.    WP02
2
 and  WP03,
3
 on  detection  
technology  review  and  the  ethical  norms  of  counter-­‐terrorism  respectively,  
developed  a  framework  of  ethical  analysis  that  serves  as  a  useful  basis  for  
1
2
 http://detecter.eu/    
 See,  for  example:  D12.2.10  ‘Detection  Technology  Quarterly  Update  10’  
www.detecter.bham.ac.uk/pdfs/D12_2_10_QuarterlyUpdateonTechnology_10__1_.doc    
3
 See  in  particular:  D5.1  ‘The  Moral  Risks  of  Preventive  Policing’  
http://www.detecter.bham.ac.uk/pdfs/D05.1MoralRisksofPreventivePolicingv2.pdf,    
D5.2  ‘The  Relative  Moral  Risks  of  Detection  Technology’  
www.detecter.bham.ac.uk/pdfs/D05.2.The_Relative_Moral_Risks_of_Detection_Technology.doc    and    
D5.3.  ‘Taking  Moral  Risks  Given  an  Analysis  of  what’s  Wrong  with  Terrorism’  
www.detecter.bham.ac.uk/pdfs/D05.3.TakingMoralRisksv2.doc    
 
2  
considering  the  ethical  norms  of  use  of  surveillance  technology  in  serious  crime  
more  generally.    In  section  1  I  outline  this  framework.  
 
The  DETECTER  project  identified  three  distinct  categories  of  harm  of  detection  
technologies:  intrusion,  error  and  damage  to  trust.    Intrusion  is  understood  in  terms  
of  penetration  of  a  normatively  protected  zone  of  a  person  or  their  life.    Normatively  
protected  zones  of  privacy  are  breached  by  looking  uninvited  into  a  changing  room,  
or  by  looking  uninvited  through  somebody’s  correspondence.  
 
At  least  three  categories  of  normative  protection  associated  with  the  concept  of  
privacy  can  be  identified.    This  is  normative  in  the  same  sense  that  there  are  
normatively  sustained  conventions  against  lying  –  normative  protections  in  this  
sense  are  quite  distinct  from  legal  protections.    The  norms  of  privacy  in  question  
include  the  following:  respect  for  bodily  privacy,  particularly  the  privacy  of  the  naked  
body;  respect  for  privacy  of  home  spaces;  and  finally  respect  for  private  life  –  
matters  of  conscience  and  association  understood  to  be  private  matters  even  when  
pursued  in  public  places  such  as  places  of  worship  or  libraries.  
 
Surveillance  technologies  may  intrude  on  bodily  privacy  when  they  scan  the  body  
directly,  as  is  the  case  with  certain  radar  scanners  and  millimetre  wave  full  body  
scanners.    Bodily  privacy  may  also  be  intruded  upon  by  video  or  audio  technologies  if  
they  are  placed  in  areas  such  as  changing  rooms  which  are  widely  understood  as  
being  protected  from  observation.  
 
Likewise,  homes  are  widely  understood  as  protected  from  others’  observation.    The  
home  is  the  place  that  one  has  greatest  latitude  to  do  as  one  pleases  without  the  
scrutiny  or  interference  of  others.    Hotel  rooms  can  take  on  a  similar  (albeit  
temporary)  significance  for  a  guest  occupying  them,  and  thus  bugs  or  miniaturised  
cameras  placed  in  such  places  can  be  highly  intrusive  in  the  same  way  as  if  they  were  
placed  in  the  home.  
 
 
3  
We  additionally  have  a  concept  of  ‘private  life’  that  covers  much  of  the  life  that  is  led  
outside  the  home;  for  example,  when  one  arranges  to  go  to  a  restaurant  with  a  
romantic  partner  or  attend  a  meeting  of  a  local  religious  organisation  in  a  place  of  
worship  explicitly  open  to  all,  such  activities  may  reasonably  be  thought  part  of  one’s  
private  life.    This  is  a  weaker  form  of  privacy,  and  cannot  rule  out  all  observation  –  
after  all  one  might  not  be  able  to  help  seeing  a  couple  at  dinner  in  a  restaurant  if  one  
is  dining  there  oneself.    However,  it  does  rule  out  persistent  attention,  
eavesdropping  or  following  as  intrusive  behaviour  in  need  of  justification.    
Technologies  can  penetrate  the  privacy  of  private  life  by  virtue  of  their  ability  to  
track  an  individual’s  movements  and  activity.    Furthermore,  bugging  and  telephone  
taps  are  intrusive  in  part  because  of  what  they  reveal  about  the  individual’s  private  
life.  
 
Intrusion  is  not  the  only  significant  ethical  risk  associated  with  detection  and  
surveillance  technologies.    Errors  may  be  harmful  when  they  lead  to  false  arrest  or  
harassment.    The  most  extreme  consequences  of  error,  such  as  miscarriage  of  
justice,  are  arguably  even  more  significant  than  the  most  extreme  intrusions.    
However,  the  intention  is  not  comparison  between  the  different  categories  of  risk.    
The  framework  is  intended  to  identify  the  different  kinds  of  ethical  danger  that  
determine  the  overall  riskiness  of  different  techniques,  technologies  and  
investigations.    Investigations  invariably  pose  some  risk  of  error,  of  false  suspicion  
and  inconvenience  to  innocent  people.    However,  certain  kinds  of  investigation  and  
especially,  for  example,  those  in  preventive  counter-­‐terrorism    are  particularly  prone  
to  the  false  identification  of  suspects,  because  there  is  often  very  little  evidence  to  
rely  upon.
4
   
 
4
 Which  may  well  combine  disastrously  with  a  high  public  demand  for  prosecution  –  see,  for  
example:(Adam  Roberts,  1989,  60)  “Its  main  problems  arise  from  the  fact  that  it  involves  trying  to  
combat  clandestine  fighters,  who  may  cause  the  most  appalling  carnage,  but  who  hide  among  the  
rest  of  the  population  and  are  very  difficult  to  track  down.    This  creates  a  situation  where  there  is  
often  a  strong  public  desire  for  retribution,  but  the  proper  target  for  such  retribution  is  not  available.”  
 
4  
There  has  been  much  public  coverage  of  databases  of  existing  suspects  and  data  
mining  programmes  used  to  identify  terrorist  suspects.
5
   Much  public  criticism  of  
these  techniques  has  called  attention  to  their  intrusiveness,
6
 but  the  large  scope  for  
error  seems  to  be  the  matter  of  greater  concern.    Both  databases  and  data  mining  
may  be  error  prone  due  to  problems  of  name  matching  (identifying  intelligence  in  a  
database  with  a  named  individual),
7
 and  data  mining  techniques  often  generate  
many  false  matches,  particularly  in  the  counter-­‐terrorism  context.
8
 
 
The  injustice  of  discrimination  overlaps  with  the  moral  risk  of  error,  as  it  can  be  both  
a  cause  and  an  effect  of  error.    It  is  a  cause  of  error  if  an  individual  incorrectly  
identifies  someone  as  a  suspect  due  to  their  own  discrimination.    Discrimination  may  
also  be  an  effect  of  error  if  error  resulting  from  a  technical  or  management  process  
systematically  casts  suspicion  on  a  particular  category  of  person.    For  example,  a  
number  of  smart  camera  systems  trigger  alerts  at  what  is  categorised  as  ‘abnormal  
activity’
9
 –  if  this  systematically  identifies  innocuous  activity  on  the  part  of  a  
particular  ethnic  minority  as  ‘abnormal’,  and  they  are  repeatedly  stopped  and  
questioned  as  a  result,  then  this  is  discriminatory.      
5
 See,  for  example:  http://www.nytimes.com/2008/10/08/washington/08data.html?_r=0  and  
http://www.guardian.co.uk/uk/2009/feb/25/database-­‐state-­‐ippr-­‐paper,  and  
http://www.aclu.org/technology-­‐and-­‐liberty/feature-­‐capps-­‐ii.    
6
7
 See,  for  example:  Tavani,  1999.    
 See  for  example  the  DETECTER  Deliverable  D5.2  “Misspellings,  spelling  variations  among  
phonetically  identical  names  (e.g.  Jeff  and  Geoff),  the  lack  of  any  standard  representation  of  names  
from  a  number  of  languages  that  do  not  use  the  Roman  alphabet,  the  use  of  nick  names,  titles,  
permutations,  abbreviations  and  omissions  of  names  (which  vary  by  culture),  the  use  of  definite  
descriptions  (e.g.  ‘the  Prime  Minister  of  Great  Britain’  vs.  ‘Tony  Blair’)  and  name  changes  over  time  all  
provide  sources  of  error  which  may  result  in  unjust  sanction”  and  Branting,  L.  Karl.  2005,  ‘Name  
Matching  in  Law  Enforcement  and  Counter-­‐Terrorism’  
8
 As,  for  example,  notoriously  with  the  German  ‘Rasterfahndung’,  identifying  suspects  by  having  come  
from  an  Islamic  country,  ‘being  registered  as  a  student’,  and  being  a  male  between  18  and  40  years  of  
age.    The  system  identified  300,  000  individuals,  and  resulted  in  no  arrests  or  prosecutions.    On  a  
range  of  other  counter-­‐terrorism  data  mining  programmes  see  DETECTER  Deliverable  D8.1.  
www.detecter.bham.ac.uk/pdfs/D8.1CounterTerrorismDataMining.doc    
9
 See,  for  example:  Behavioural  Recognition  Systems’  AIsight  2.1  
http://www.brslabs.com/files/pdf/AISight_2%201_Final.pdf.  
 
5  

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin